It is time for #CBDC (Central Bank Digital Currency) reference architecture

1 Executive summary

CBDC implementation is a daunting task and its the most difficult feature is legality through CBDC life cycle. Let us architect CBDC as commonly agreed reference architecture which is available to everyone to move forward further (via improving the entire global economic position) and faster (via innovations).

Although there are some attempts to outline CBDC architectures these attempts are still missing good methodology practices.

2   Introduction

Leading financial and economic organisations as the BIS, WEF and FATF define CBDC as digital representation of cash, bearer financial asset and legal tender. In other words CBDC is “money M0 in digital form”.

For several reasons, the interest for CBDC has considerably increased during the year 2020 – there are a lot of papers, some central banks develop CBDC requirements, some central banks announced (sometimes mistakenly) that they are already implemented CBDC and some banks invite professional community to develop CBDC solutions. International financial organisations also actively participate and publish their requirements (e.g. the BIS at a WEF conference on July 2020 requested systemic stability for the Financial Market Infrastructure).

As everyone knows that Digital Transformation (DT) removes many of artificial and natural barriers. DT of money is about moving physical cash (the most basic and popular form of money) into digital space. Such digital representation of cash or digital cash or CBDC will have almost the same functionality as physical cash without barriers of physical representation.

Digital cash (which is done correctly) reduces many of existing barriers: everyone may execute legally and free-of-charge digital payments domestically and internationally. This is a huge difference with the current situation when a lot of people have no bank accounts and each cross-border payment costs approximately $ 35 (another estimation is 7-10% of remittance amount).

Digital cash is an enabler for frictionless, instantaneous, unconditional transfer with legal finality of value between any two parties; anytime, anywhere, without censorship. This will enable DT of financial market and investment practices. Lesser barriers in time, distance, legality, etc. improve the entire global economic position of all of the earths peoples.

The representation of money (physical or digital) is irrelevant, the sole characteristic required of any money based medium of exchange, is to achieve global unconditional settlement with legal finality, or legal tender based absolute cancellation of a payment obligation. Any digital bearer asset which is accepted by the community, to satisfy both of the two functions: a medium of exchange and a unit of account; with guaranteed unconditional settlement and achieve legal finality is Cash or Legal Tender. 

3 Why reference architecture

The FinTech and crypto communities already admitted that CBDC is not easy. Many people said that CBDC will take years to implement because of CBDC implementation complexity. Some systemic factors of complexity are listed below:
  • Each country is different.
  • There are many similarities between countries.
  • Even Central Banks are confused with CBDC (see LBCOIN case).
  • National CBDC is legal only in its own country.
  • CBDC must be legal in each of its action.
  • Least changes in legislation (see Russian Parliament “DUMA” case).
  • To enable cross-border payments via CBDC an additional mechanism will be necessary.
  • A systemic approach is needed to guarantee stability of operations (BIS said this at a WEF conference on July 2020).
  • CBDC must enable innovations.
  • CBDC must enable public-private partnership.
  • CBDC must prevent the Wirecard case ([1] and [2]).
  • etc.
Thus it is mandatory to develop about 200 rather similar CBDC implementations which possess the same emerging features. This is the case for the pattern “reference architecture”. This pattern is about two principal steps and the optional third one:
  1. developing an agreed abstract (or imaginary) architecture which guarantees desired emerging features and
  2. using such a reference architecture as a template for solution architecture of each CBDC implementation.
  3. making a reference implementation.

A reference architecture is universally valid within a particular problem space or subject field. A reference architecture is technology-neutral. An important driving factor for the creation of a reference architecture is to improve the effectiveness of creating products, product lines and product portfolios by
  • managing synergy (e.g. to agree on a common infrastructure),
  • providing guidance, e.g. architecture principles and good practices,
  • providing an architecture baseline and an architecture blueprint, and
  • capturing and sharing (architectural) patterns.
In this short article it is not possible to cover all the complexity of the CBDC reference architecture. Therefore, only factor “CBDC must be legal in each of its action” will be discussed below. 

4 CBDC must be legal in each of its action

Each CBDC implementation must implement all legal features of cash and guarantee integrity of CBDC through its life cycle. CBDC life cycle is rather simple and only the difficult part is payments:

1) Emission by the Central Bank or an operator on behalf of the Central Bank

2) Participation in payments between buyers and sellers

3) Recuperation CBDC by the Central Bank by taxation process and potential reuse of CBDC.

The status CBDC as “legal tender” implies that CBDC payments are protected by the State to take place with sufficient certainty to satisfy the needs of buyer sand sellers. Buyers must be protected by ensuring that payment employing CBDC may not be refused. Sellers will accept CBDC for its value only if they can be certain that it will, in turn, be accepted from them for the same value.

There are many national laws and other governmental instructions which regulate the current practices of cash usage. It is not necessary to know all of laws and instructions but the current practices must be carefully reproduced or reengineered in the digital space.

In the CBDC retail scenario, everyone may have one or many “heaps” of cash. To pay some cash, a payer must collect the necessary amount of cash from all his/her heaps of cash and pass this heap of cash to a payee. The latter validates this heap of cash and, usually, accepts it. The payment is done with its legal finality. Thus in digital space all of such actions may be explicit (because easy to check and guarantee legality of each action) and be executed as one transaction “all or nothing”.

In the CBDC wholesale scenario, if CBDC is guaranteed being legal then RTGS for CBDC is not much different from the CBDC retail scenario described in the previous paragraph. Of course, it must be remembered that such RTGS takes place in the same country (national CBDC is legal only in its own country) and CBDC is not mixed with commercial bank money which are not legal tender (please note that IMF’s synthetic CBDC and Swedish e-krona are such mixtures). More interesting, if CBDC wholesale can be carried out without the Central Bank then there is no need for RTGS for CBDC. It is a very useful simplification for the national financial market infrastructure.

In the CBDC cross-border payments scenario, an additional Universal CBDC (U-CBDC) must be internationally (or regionally) created and legalised. Let us consider that nation AAA CBDC (CBDC-AAA) must used to pay in nation BBB CBDC (CBDC-BBB). The price of the payment is expressed in U-CBDC and agreed by payer and payee. Then the sequence of actions is the following:
  1. Payer asks the country AAA Central Bank (CB-AAA) to buy agreed amount of U-CBDC.
  2. The CB-AAA buy on a free market that U-CBDC with CBDC-AAA.
  3. Payer transfers that U-CBDC to payee.
  4. Payee asks the country BBB Central Bank (CB-BBB) to sell on a free market that U-CBDC for CBDC-BBB.
All these actions must be legal and the whole sequence must be “all or nothing”, ideally, as requested by BIS DvP 1 model.
And, another big challenge in the CBDC cross-border case will be the BIS principles and CPSS-IOSCO principles for financial market infrastructures[3]. For example:

“Principle 1: Legal basis

An FMI should have a well-founded, clear, transparent, and enforceable legal basis for each material aspect of its activities in all relevant jurisdictions.” 

5 Examples of weak methodology

Reading of dozens various articles about CDBC, several cases of “broken logic” were identified. These cases are analysed below from international practices for developing reference architectures. Good, right and successful reference architectures must be based on an agreed set of concepts, stakeholders, stakeholder’s concerns, use cases, requirements, desired characteristics of potential solutions, architecture principles (decisions) for potential solutions, etc. Some of them belong to the problem space (which is solution-independent) and some of them belong to the solution space (which defines a set of potential solutions).

Usually, reference architectures are technology-neutral to promote innovations. Reference architectures make all the decisions explicit and explainable.

The IMF[4] is one of the first financial institutions which actively leads CBDC discussions. However, their logic is not always faultless. An example. The IMF impose in solution architectures the mandatory involvement of private companies, however, this architecture decision for potential solutions (from the solution space) is not explicitly justified and linked with the CBDC problem space description. Thus this architecture decision is like "putting a chariot before the horse".

Another example. IMF's synthetic CBDC (sCBDC) “ is a liability of private firms—the sCBDC issuers—rather than of the central bank”. Thus sCBDC is not legal tender and it can’t be compared with CBDC. Such kind of error could occur because of the absence of well-developed CBDC reference architecture.

Germany, federal finance ministry published a very good report[5]. However, their architecture decision for potential solutions to fix “distributed ledger technology (DLT) or blockchain technology” for CBDC is not logical. Because the solution space may have several possible solutions, such fixing is an artificial and unexplainable self-limitation. It stopped innovations. A better logic would be keeping the CBDC reference architecture technology-neutral.

Lithuania Central Bank “The LBCOIN is very similar to what is known as a central bank digital currency, said Jurgilas, putting Lithuania at the forefront of development of a fiat digital currencies.” [6]

However, their LBCOIN is not a legal tender hence not CBDC thus LBCOIN is , maximum, a stablecoin. Such an error can be found via a CBDC reference architecture.

Former CFTC Commissioner[7]: “Bowen made it very clear that even though the federal reserve chairman is a big believer that the government should develop this digital currency, bringing in multiple stakeholders will give a diverse perspective on its development and could give a better result to the overall project.”

Careful identification of stakeholders is a mandatory step for developing a good reference architecture.

Russian parliament Duma[8]: After 2 years of delays have approved as a law something strange which is not necessary aligned with FATF standards.

Careful check of all international requirements is a mandatory step for developing a good reference architecture.

ITU-T Focus Group Digital Currency group Digital Fiat Currency report[9] is the most comprehensive documentation about CBDC reference architectures. For the year 2019 is was perfect, however, several points are slightly different in the year 2020. Namely, terminology about CBDC.

This reference architecture doesn’t provide enough information about the problem space thus some of architecture decisions are not traced from use needs. It seems that the majority of architecture decisions are grew bottom-up which is not usual practice for developing reference architectures. As the result, this reference architecture is more an overview of existing approaches (e.g. retail CBDC, wholesale CBDC and cross-border payments CBDC are different) and technology-driven as confirmed by some considerations below:

“Another operational case in point for such smart contracts include DLT-based liquidity management facilities, which would enable bilateral transaction netting.”

“DFC using DLT architecture will be as much successful as the maturity of the ecosystem that is ready to embrace its full potential.”

This reference architecture looks like technology architecture. 

6 Conclusion

Because of the world-wide importance of money, there is a serious demand for a commonly agreed CBDC reference architecture. This reference architecture must be created with good methodology practices.

This ITU initiate has been recently relaunched as “Digital Currency Global Initiative” (https://www.itu.int/en/ITU-T/extcoop/dcgi/Pages/default.aspx) – hope they will use good methodology practices.

7 Additional resources

Publicly available documents.










[1] https://www.coindesk.com/from-enron-to-wirecard-how-blockchain-tech-could-have-helped

[2] https://www.linkedin.com/posts/wassim-jarkas_from-enron-to-wirecard-how-blockchain-tech-activity-6693044727941488640-UPxs

[3] https://www.bis.org/cpmi/publ/d101a.pdf

[4] https://www.coindesk.com/private-sector-could-bring-value-to-future-cbdc-launches-says-imf-official

[5] https://www.bundesfinanzministerium.de/Content/DE/Downloads/Finanzmarktpolitik/2020-07-08-fintechrat-digitaler-euro.pdf?__blob=publicationFile&v=3

[6] https://www.reuters.com/article/us-eu-cryptocurrency-lithuania/lithuania-dabbles-in-https://cryptodaily.co.uk/crypto-coin-as-central-banks-look-for-ways-to-fend-off-facebook-idUSKBN2431RF

[7] https://cryptodaily.co.uk/2020/07/former-cftc-us-cbdc

[8] https://www.rbc.ru/crypto/news/5f16c6379a794732b6dd31e7

[9] https://www.itu.int/en/ITU-T/focusgroups/dfc/Pages/default.aspx